vgvassilev wrote:
> @alexfh thank you very much!
>
> @vgvassilev but we have to provide similar mechanism, so it is allowed to get
> all the specializations for a templated decl.
I think this is the code snippet we are bitten by:
https://github.com/llvm/llvm-project/blob/d1fae5996e66c2a9f0b1c
ChuanqiXu9 wrote:
@alexfh thank you very much!
@vgvassilev but we have to provide similar mechanism, so it is allowed to get
all the specializations for a templated decl.
https://github.com/llvm/llvm-project/pull/83237
___
llvm-branch-commits mailing
tblah wrote:
> They won't be able to that regardless. The OpenMP library will have to match
> what Flang is using; otherwise, two OpenMP implementations will co-exist in
> the same process and that's usually a recipe for problems.
I probably wasn't clear. I meant something like `-fopenmp=libom
Author: Shubham Sandeep Rastogi
Date: 2024-10-23T14:30:35-07:00
New Revision: 565b370162888c91e82bd5c9a8787bbc4942219c
URL:
https://github.com/llvm/llvm-project/commit/565b370162888c91e82bd5c9a8787bbc4942219c
DIFF:
https://github.com/llvm/llvm-project/commit/565b370162888c91e82bd5c9a8787bbc4942
@@ -69,6 +81,241 @@ FunctionPass *llvm::createAMDGPURBLegalizePass() {
using namespace AMDGPU;
+const RegBankLegalizeRules &getRules(const GCNSubtarget &ST,
+ MachineRegisterInfo &MRI) {
+ static std::mutex GlobalMutex;
+ static SmallDens
@@ -54,9 +54,28 @@ const MachineBasicBlock
*MachineSSAContext::getDefBlock(Register value) const {
return F->getRegInfo().getVRegDef(value)->getParent();
}
+static bool isUndef(const MachineInstr &MI) {
+ return MI.getOpcode() == TargetOpcode::G_IMPLICIT_DEF ||
+ M
https://github.com/petar-avramovic edited
https://github.com/llvm/llvm-project/pull/112882
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/petar-avramovic edited
https://github.com/llvm/llvm-project/pull/112882
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -0,0 +1,334 @@
+//===-- AMDGPURBLegalizeRules.cpp
-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
@@ -69,6 +81,241 @@ FunctionPass *llvm::createAMDGPURBLegalizePass() {
using namespace AMDGPU;
+const RegBankLegalizeRules &getRules(const GCNSubtarget &ST,
+ MachineRegisterInfo &MRI) {
+ static std::mutex GlobalMutex;
+ static SmallDens
@@ -69,6 +81,241 @@ FunctionPass *llvm::createAMDGPURBLegalizePass() {
using namespace AMDGPU;
+const RegBankLegalizeRules &getRules(const GCNSubtarget &ST,
+ MachineRegisterInfo &MRI) {
+ static std::mutex GlobalMutex;
+ static SmallDens
@@ -107,3 +107,183 @@ void IntrinsicLaneMaskAnalyzer::findLCSSAPhi(Register
Reg) {
S32S64LaneMask.insert(LCSSAPhi.getOperand(0).getReg());
}
}
+
+MachineInstrBuilder AMDGPU::buildReadAnyLaneB32(MachineIRBuilder &B,
+const
@@ -0,0 +1,334 @@
+//===-- AMDGPURBLegalizeRules.cpp
-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
@@ -69,6 +81,241 @@ FunctionPass *llvm::createAMDGPURBLegalizePass() {
using namespace AMDGPU;
+const RegBankLegalizeRules &getRules(const GCNSubtarget &ST,
+ MachineRegisterInfo &MRI) {
+ static std::mutex GlobalMutex;
+ static SmallDens
@@ -227,11 +227,11 @@ void createHLFIRToFIRPassPipeline(mlir::PassManager &pm,
bool enableOpenMP,
hlfir::createOptimizedBufferization);
}
pm.addPass(hlfir::createLowerHLFIROrderedAssignments());
+ if (enableOpenMP)
+pm.addPass(
https://github.com/petar-avramovic edited
https://github.com/llvm/llvm-project/pull/112864
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/petar-avramovic edited
https://github.com/llvm/llvm-project/pull/112864
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/petar-avramovic edited
https://github.com/llvm/llvm-project/pull/112863
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -63,4 +70,189 @@ char &llvm::AMDGPURBSelectID = AMDGPURBSelect::ID;
FunctionPass *llvm::createAMDGPURBSelectPass() { return new AMDGPURBSelect(); }
-bool AMDGPURBSelect::runOnMachineFunction(MachineFunction &MF) { return true; }
+bool shouldRBSelect(MachineInstr &MI) {
+
@@ -63,4 +70,189 @@ char &llvm::AMDGPURBSelectID = AMDGPURBSelect::ID;
FunctionPass *llvm::createAMDGPURBSelectPass() { return new AMDGPURBSelect(); }
-bool AMDGPURBSelect::runOnMachineFunction(MachineFunction &MF) { return true; }
+bool shouldRBSelect(MachineInstr &MI) {
+
https://github.com/petar-avramovic edited
https://github.com/llvm/llvm-project/pull/112863
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/ivanradanov edited
https://github.com/llvm/llvm-project/pull/113082
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
mjklemm wrote:
> * The above requirement basically breaks allowing users to pick their own
> openmp library when compiling their application. I'm not sure how widely used
> this is.
They won't be able to that regardless. The OpenMP library will have to match
what Flang is using; otherwise, t
tblah wrote:
> How do we usually have convergence on the definition of these runtime
> functions when we introduce a library? Is it usually through a RFC? Because
> if we think of vendors swapping implementations, does there need to be a
> consensus prior to the implementation?
To be clear, I
vgvassilev wrote:
I think we almost never should call findAll. Most of the time we have a
concrete template argument list…
https://github.com/llvm/llvm-project/pull/83237
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://
Author: Lei Huang
Date: 2024-10-23T09:35:02-04:00
New Revision: 84d44efd214f0f7be39e7e063b2d94344c3ff979
URL:
https://github.com/llvm/llvm-project/commit/84d44efd214f0f7be39e7e063b2d94344c3ff979
DIFF:
https://github.com/llvm/llvm-project/commit/84d44efd214f0f7be39e7e063b2d94344c3ff979.diff
LOG
@@ -4,6 +4,7 @@
// https://github.com/llvm/llvm-project/issues/47137
#include
#include
+#include
cjappl wrote:
Cool, thanks for the confirmation, resolving this comment
https://github.com/llvm/llvm-project/pull/96507
___
alexfh wrote:
BTW, if you wonder about the callers of `LoadExternalSpecializations`:
```
- 90.30% 0.00% clang clang [.]
clang::Sema::CheckVarTemplateId(clang::VarTemplateDecl*, clang::SourceLocation,
clang::SourceLocation, clang::TemplateArgumentListInfo const&)
alexfh wrote:
Compiling Clang with more debug information and without FDO+ThinLTO indeed
shows
`clang::serialization::MultiOnDiskHashTable::findAll()`
as the main time sink:
```
+ 96.99% 0.00% clangclang [.]
clang::Sema::runWithSufficientStackSpace(clang::SourceLocati
NimishMishra wrote:
Thanks for the inputs @tblah.
>> Having it in a runtime library would make it easier for vendors to swap out
>> their own implementations e.g. maybe somebody already has a great MATMUL for
>> AMD GPUs and wants flang to use that, but the implementation is useless on a
>>
tblah wrote:
> * I imagine that using a C `omp for` inside of a fortran `omp parallel`
> should work **so long as they use the same openmp library**. Of course care
> will need to be taken about privatisation etc but that can probably be
> handled in the interface of the runtime function.
@@ -335,49 +336,129 @@ static void parallelizeRegion(Region &sourceRegion,
Region &targetRegion,
for (auto [i, opOrSingle] : llvm::enumerate(regions)) {
bool isLast = i + 1 == regions.size();
if (std::holds_alternative(opOrSingle)) {
-OpBuilder singleBu
@@ -335,49 +336,129 @@ static void parallelizeRegion(Region &sourceRegion,
Region &targetRegion,
for (auto [i, opOrSingle] : llvm::enumerate(regions)) {
bool isLast = i + 1 == regions.size();
if (std::holds_alternative(opOrSingle)) {
-OpBuilder singleBu
@@ -335,49 +336,129 @@ static void parallelizeRegion(Region &sourceRegion,
Region &targetRegion,
for (auto [i, opOrSingle] : llvm::enumerate(regions)) {
bool isLast = i + 1 == regions.size();
if (std::holds_alternative(opOrSingle)) {
-OpBuilder singleBu
@@ -227,11 +227,11 @@ void createHLFIRToFIRPassPipeline(mlir::PassManager &pm,
bool enableOpenMP,
hlfir::createOptimizedBufferization);
}
pm.addPass(hlfir::createLowerHLFIROrderedAssignments());
+ if (enableOpenMP)
+pm.addPass(
https://github.com/tblah edited https://github.com/llvm/llvm-project/pull/113082
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/tblah commented:
I don't have a strong opinion about whether this should be done in flang
codegen or in a runtime library. Some drive-by thoughts:
- I imagine that using a C `omp for` inside of a fortran `omp parallel` should
work **so long as they use the same openmp librar
https://github.com/Thirumalai-Shaktivel ready_for_review
https://github.com/llvm/llvm-project/pull/113082
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -4,6 +4,7 @@
// https://github.com/llvm/llvm-project/issues/47137
#include
#include
+#include
gbMattN wrote:
The test uses uint64_t, which is in stdint.h. When I was running these tests,
not having that file included gave me a compile error.
https://gi
https://github.com/kovdan01 approved this pull request.
LGTM, thanks!
https://github.com/llvm/llvm-project/pull/110705
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-c
https://github.com/kovdan01 edited
https://github.com/llvm/llvm-project/pull/110705
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -1757,38 +1762,70 @@ void
AArch64AsmPrinter::emitPtrauthCheckAuthenticatedValue(
//Lsuccess:
// ...
//
- // This sequence is expensive, but we need more information to be able to
- // do better.
- //
- // We can't TBZ the poison bit because EnhancedPAC2 X
42 matches
Mail list logo
